When does a change I make take effect; does it go live as soon as I save it?

Short answer

Yes. When Save, Add or Submit is pressed, the save is accepted immediately; Load Balancer changes are held for about one second, merged and applied live. The progress of applying the change is followed as a percentage on the sync indicator at the bottom right of the screen (Syncing Traffic manager...); when the process ends, the indicator shows All synced. In HTTP and L7 TCP vServices, the change is mostly applied without open connections being dropped; only a few settings restart the vService. In Network vServices, every save restarts the vService. In a cluster setup, with the default auto-sync, the change is also sent to the peer device automatically; if Manual Sync Mode is on, the change stays on the local device and is sent manually with Synchronize.

How Does It Work?

1. Save: The change is saved and the screen returns immediately.
2. Short wait: For Load Balancer changes, TR7 waits one second; other changes made during this time are merged into a single application.
3. Application: The new configuration is generated and applied only to the vServices that changed; vServices that did not change are not touched.
4. Order: If new changes arrive while an application is in progress, they are put on hold and, when it ends, applied at once in their most up-to-date state; no change is lost.

Where Is the Application Monitored?

The sync indicator at the bottom right corner of the screen shows the application of the saved change to the system live:

Indicator	Meaning
Syncing Traffic manager... and a percentage bar next to it	The Load Balancer configuration is being applied; the percentage shows how much of the application has been completed. In network changes, the field name is shown as Network.
Syncing Traffic manager (both devices)...	In the cluster, the change is being applied on both devices at once.
All synced (green check)	The application is complete; all fields are up to date.
<N> sync error(s) (red)	The application gave an error; when the pointer is hovered over the indicator, it is listed in which field, on which device and with which error it stopped. The error stays visible until that field is next applied successfully.
<N> field(s) out of sync	In the cluster, the configurations of the two devices differ (e.g. changes made while Manual Sync Mode is on).

The Active Operations window that opens when the pointer is hovered over the indicator shows the percentage of each operation and the device on which it runs; when the indicator is clicked, the HA Clustering & Sync screen opens. The same summary also appears in the Synchronization Status row on the Home screen.

Are Open Connections Dropped?

Change	Effect
Ordinary changes in an HTTP / L7 TCP vService (Backend Service, action, condition, certificate, profile…)	Seamless reload; open connections are preserved
Process-level settings listed in the table below in an HTTP / L7 TCP vService	The vService is restarted; open connections are dropped
Any change in a Network vService (including a name change)	The vService is restarted
Right-click → Reload in the vService list	The vService is forcibly restarted; open connections are dropped

If the seamless reload fails, TR7 still applies the new configuration by restarting the vService.

Settings That Restart an HTTP / L7 TCP vService

The following fields cannot be applied with a seamless reload; when their values change, the vService is restarted during the save:

Where	Field
vService → Details → QOS	Cores, Memory Limit, CPU Priority, Cpu Usage Limit (%) and Core Engine
vService Limits profile	Max. Connections/s, Max. Session Rate/s, Max. Total SSL Connections, Max. SSL Connections/s and Buffer Size(kB)
WAF (HTTP only)	Turning WAF on or off in the vService, and Max. Body Size to Process(kB) in the WAF options
Compression (HTTP only)	Turning Compression on or off in the vService (Details → Compression), and Compression Level in the profile
HTTP 1.0 Header Compatibility action	Adding the action to or removing it from the vService, and changing the Headers to Fix list

• Profile fields change in two ways: when the profile assigned to the vService is edited, or when another profile is selected in the vService. If the value of a field in the table changes, both restart the related vServices.
• When QOS is turned off, Cores, Memory limit, CPU priority and Cpu Usage Limit (%) return to their defaults; if there is a value different from the default in these fields, turning off QOS also restarts the vService.
• The buffer size of the vService is the larger of the Buffer Size(kB) and (if WAF is on) Max. Body Size to Process(kB) values. The restart happens when this larger value changes (this is also why turning WAF on and off restarts it); for example, in a vService with WAF on, changing Buffer Size(kB) does not restart the vService as long as it stays below the WAF value.

In a Cluster

The synchronization method in the cluster is determined by the Manual Sync Mode field on the Settings → Device Management → HA Clustering & Sync screen:

Mode	Behavior
Auto-sync (default; Manual Sync Mode off)	While the change is applied on the local device, it is also sent to the peer device automatically. Load Balancer changes pass to the peer device in merged form after the one-second wait on the local device ends.
Manual Sync Mode on	The change is applied only on the local device; the indicator at the bottom right shows <N> field(s) out of sync. The changes are sent to the peer device with Synchronize on the HA Clustering & Sync screen; there must be at least 30 seconds between two Synchronize operations.

Common Misconceptions

• "For the change to take effect, I must Reload the vService." It is not required; the save is applied on its own. Reload restarts the vService and drops the open connections.
• "Every change reloads all the vServices." It is applied only to the vServices that the change affects.
• "Changing the name in a Network vService is harmless." In a Network vService, every save restarts the vService; the change is made at a low-traffic time.
• "While Manual Sync Mode is on, the change is also applied to the peer device." It is not applied; until Synchronize is pressed, the peer device runs with the old configuration.

How it is done in the documentation

• HA Clustering & Sync · Manual Sync Mode
• HA Clustering & Sync · The Sync Summary Table
• Summary · Device info (Synchronization status)
• Network vService · Save
• vService · Right-Click Menu

Related questions

• What does Reload in the right-click menu do; how is it different from Disable / Enable?
• Can I restart a vService without interruption (soft / hard restart)?
• Is the configuration automatically synchronized between cluster members?
• The change I made does not seem to take effect; what should I check?